Economic Research Southern Africa (ERSA) is a research programme funded by the National

Treasury of South Africa. The views expressed are those of the author(s) and do not necessarily represent those of the funder, ERSA or the author’s affiliated

institution(s). ERSA shall not be liable to any person for inaccurate information or opinions contained herein.

The Impact of the COMESA-EAC-SADC

Tripartite Free Trade Agreement on the

South African Economy

L. Walters, H.R. Bohlmann and M. W. Clance

ERSA working paper 635

September 2016

The Impact of the COMESA-EAC-SADCTripartite Free Trade Agreement on the South

African Economy

L. Walters, H.R. Bohlmann and M.W. Clance

September 13, 2016

Abstract

This paper analyses the e¤ects of the COMESA-EAC-SADC TripartiteFree Trade Agreement (TFTA) on the South African economy using aglobal Computable General Equilibrium (CGE) model. Simulation resultsshow that South Africa’s economy gains from the implementation of thetrade agreement with GDP rising by more than 1 per cent relative tothe baseline. This win in overall economic activity occurs on the back ofa terms of trade increase and a surge in regional trade, which allows forhigher levels of both exports and imports. The boost to exports stimulateslocal industries, whilst relatively cheaper imports lead to welfare gainsfor local consumers. Increased trade and industry activity causes higherdemand for endowments, including skilled and unskilled labour, capitaland land, pushing up wages and capital rentals.

JEL Codes: C68, F13, O55Keywords: Computable General Equilibrium (CGE) Modelling, Free

Trade Agreement, South Africa

1 Introduction

The National Development Plan for 2030 clearly states the importance of in-creased international relations and trade for South Africa’s economic future.“. . . its policies on African integration must be based on positioning South Africaas one of the continent’s powerhouses that would lead African development andin‡uence world a¤airs” (NPC, 2011). The COMESA-EAC-SADC TripartiteFree Trade Agreement (TFTA) adds to this vision for increased global partici-pation, super…cially relating to economic cooperation between South Africa andthe rest of the TFTA participants. To attain sustainable economic growth, it isimportant to set comprehensive trade agreements in place.

In this study, we investigate the implications of the TFTA on the SouthAfrican economy using a global Computable General Equilibrium (CGE) model.We analyse the impact of this policy using the standard Global Trade Analy-sis Project (GTAP) model alongside version 8.1 of the GTAP database. The

1

implementation of any new trade policy has a direct and indirect e¤ect on theeconomy, and for South Africa to reap the full bene…ts of such an agreement, itis essential to estimate and understand these e¤ects. The GTAP CGE model’smulti-country and multi-sector speci…cation allows us to capture the e¤ects ofcomplicated interactions among economic actors and determine who the winnersand losers are likely to be.

Our simulation results show that South Africa’s Gross Domestic Product(GDP) and welfare increases in the short-run after the implementation of theTFTA. South African trade shifts away ever slightly from other trading regionssuch as North America and the European Union (EU), towards other countriesparticipating in the TFTA. With overall economic activity in South Africa ex-panding, interesting changes in the local market structure emerge. Firstly, theincrease in demand for South African exports boosts local industries. How-ever, ‡owing from this boost in demand is a rise in local prices, and the termsof trade, which causes substitution away from locally produced goods towardsnow relatively cheaper imports. Depending on the structure of each individ-ual industry, in particular their exposure to these trade outcomes, winners andlosers emerge. Overall, the TFTA brings about positive economic e¤ects forSouth Africa’s economy.1

The structure of the paper is as follows: Section 2 provides a brief back-ground on the TFTA, South African trade and trade between South Africa andother participating countries. Section 3 reviews the relevant trade literature.Section 4 discusses the CGE methodology and looks at the GTAP database,model and model closure. Section 5 provides a small back-of-the-envelope modelto assist in the interpretation of results. Section 6 reports selected macroeco-nomic and industry results. Finally, Section 7 concludes the study and suggestspossible policy implications.

2 Background and Trade Pro…le

The TFTA is a proposed trade alliance between three of the eight Regional Eco-nomic Communities (RECs) recognised by the African Union. The TFTA is anumbrella organisation consisting of the Common Market for Eastern and South-ern Africa (COMESA), the East Africa Community (EAC) and the SouthernAfrican Development Community (SADC). The TFTA constitutes of 26 coun-tries and started negotiations in 2005, which is expected to be implemented in2016.2

1This study does not take into consideration non-tari¤ barriers, services trade and othertrade facilitation components of the TFTA. These additional components are likely to havean impact on the actual outcome of the TFTA, decrease the measured losses and create gainsfor participating countries.

2The 26 member countries are: Angola, Botswana, Democratic Republic of the Congo(DRC), Lesotho, Madagascar, Malawi, Mauritius, Mozambique, Namibia, Seychelles, SouthAfrica, Swaziland, Tanzania, Zambia, Zimbabwe, Burundi, Kenya, Rwanda, Uganda, Co-moros, Djibouti, Egypt, Eritrea, Ethiopia, Libya and Sudan (now two countries).

2

The aim of the TFTA is to coordinate and improve regional trading arrange-ments and programmes - ultimately leading to an African Economic Communityrealisation. This includes the improvement of trade facilitation to improve the‡ow of goods, the joint preparation and execution of infrastructure programmes,as well as the free movement of businesspersons within the region (COMESA-EAC-SADC-Tripartite). In addition, the TFTA addresses the issue of overlap-ping membership between the three RECs. With 26 countries included and acombined GDP of $860 billion and population of 590 million people (dti, 2014b);the TFTA is expected to generate signi…cant growth in Africa and speci…cally,South Africa. Section 2.1 considers South African trade, where after Section 2.2provides a trade pro…le on South Africa’s trade with other participating TFTAcountries.

2.1 South African Trade

South Africa has a rich and complex international trade history. The economicsanctions set in place against South Africa during the latter part of Apartheidrestricted exports and imports, as well as economic growth depending on trade.After the new democratic regime took power in 1994 and South Africa joinedthe World Trade Organisation in 1995, South Africa experienced an expansionin foreign trade. The further moving away from protectionism and the partici-pation in Preferential Trade Agreements (PTAs) with the EU in 1999 and SADCin 2000 contributed to South Africa’s openness and increased trade (Jordaan,2011). Figure 1 depicts South African trade as a percentage of GDP over theperiod between 1960 and 2013.

Between 2008 and 2009, South African imports and exports decreased with17.4 percent and 19.5 percent, respectively, due to the …nancial crisis (SARB,2014). According to the Minister of the Department of Trade and Industry,Rob Davies, exports to the EU decreased from €22 billion (2008) to €20 billion(2012), whilst imports from this region increased over the same period from €20billion to €25 billion (Ensor, 2014). As the EU was one of South Africa’s maintrading partners, the economic shock of the …nancial crisis was signi…cant onforeign trade and would still be evident afterwards.

South Africa’s exports increased by 12.7 percent in 2013, because of higherdemand from countries such as China, Botswana, Namibia, Germany and theUnited States (US). Similarly, trade with BRIC countries3 increased notablyover the last …ve years with China as a main driver. South Africa experiencedan increase in imports from China, which widened the trade de…cit to R70 billionin 2013 (dti, 2014a). The Annual Report for 2013/2014 by the Department ofTrade and Industry states that South Africa continues to increase trade, not onlywith traditional partners but also with other emerging markets. To give somebackground on South Africa’s trade and trade agreements, Table 1 provides alist of the main trading agreements South Africa participates in.

3BRIC countries include Brazil, Russia, India and China. These countries are consideredas emerging markets.

3

2.2 Trade Pro…le

This section provides a brief pro…le on South African imports, exports andthe tari¤ structure between South Africa and other TFTA countries.4 Figure2 illustrates the composition of goods imported from other TFTA countriesinto South Africa and Figure 3 shows the composition of South African goodsexported to other TFTA countries.5

Overall, only 6.236 percent of total South African imports originate fromother TFTA countries, whereas 35.133 percent of all imports are from the EU.Light and heavy manufactured goods make up 63.225 percent of imported goodsfrom all trading partners, whereas oil and gas extraction represent 10.251 per-cent (GTAP database). South Africa mainly imports manufactured goods aswell as oil and gas extraction goods from other TFTA countries, illustrated inFigure 2.

South Africa exports most of its tradable goods to the EU (31.649 percent ofall exports), North America (12.918 percent) and other TFTA countries (15.409percent). Overall, 61.406 percent of exports are that of manufactured goods,9.093 percent other extractions, followed by transport and communication whichequals 6.296 percent. Exports to other TFTA countries are mainly manufac-tured goods (69.742 percent), as shown in Figure 3. Additional goods exportedto other TFTA countries, include processed food as well as petroleum and coalproducts (GTAP database).

Table 2 illustrates (i) the tax (tari¤) on imports of goods from South Africainto other TFTA countries and (ii) the tax on imports of goods from otherTFTA countries into South Africa. Grains and crops and other secondary in-dustry products such as light and heavy manufactured commodities, textilesand clothing, as well as petroleum and coal products are highly taxed relativeto other goods. Tari¤s imposed on imports from other TFTA countries’ arelower than tari¤s levied on South African traded commodities. Tari¤s on grainsand crops, processed foods, as well as textiles and clothing are relatively high incomparison with tari¤s levied on other traded goods from the other countriesparticipating in the TFTA.

It is important to keep the trade pro…le in mind when discussing simulationresults in Section 6. The following section considers relevant literature on FreeTrade Agreements (FTAs), the TFTA and methodologies used to analyse theimpacts thereof on economies.

3 Literature Review

The argument for FTAs date back to the mercantilist era when economists be-lieved that the world drew from a limited “pot” of resources and that the wealthof a nation depended on a favourable balance of trade. Conversely, Adam Smith

4Other TFTA countries refer to other 25 countries participating in the TFTA, excludingSouth Africa.

5Data obtained from GTAP version 8.1 database. The aggregation of regions and goods isdiscussed in Section 4.2.

4

stated in The Wealth of Nations that the “pot” is not limited and can grow overtime, provided trade between countries is unobstructed (Kishtainy et al., 2012).Many studies that analyse trade agreements use di¤erent methodologies suchas partial and general equilibrium analysis, as well as econometric approaches.Econometric techniques, including the gravity model, are suitable for ex-postanalysis, whereas CGE models are more appropriate for ex-ante analysis (Cer-nat, 2003). This section examines various econometric and CGE approachesused in FTA studies.

Factors such as the distance between countries, the similarity of economies,the remoteness to the rest of the world and the comparative advantages of trad-ing partners determine the establishment of an FTA (Baier & Bergstrand, 2004).Moreover, countries consider the economic bene…ts that participation will bringabout.6 Unilateral and multilateral liberalisation has bene…cial impacts on thesteady-state growth of all trading countries (Ben-David & Loewy, 1998). Otherstudies condoning the in‡uence of FTAs on economic growth and internationaltrade include research by Baier and Bergstrand (2007), as well as Wacziarg andWelch (2008).

Authors analyse economic e¤ects of trade agreements similar to the TFTA,using various models and techniques.

Abedini and Peridy (2014) analyse the trade e¤ects from the Greater ArabFree Trade Area (GAFTA) between fourteen Arab countries using the new grav-ity model accompanied by supply-demand export equations in an imperfect com-petition environment. To control for the problem of endogeneity, the authorsutilise instrumental Generalised Methods of Moments (GMM) panel estimation.The …ndings of this paper conclude that GAFTA countries could bene…t fromdeeper regional trade integration due to the limited bene…ts of GAFTA.

Other econometric methods include the Michigan Model of World Produc-tion and Trade employed by Brown, Kiyota and Stern (2008) to determine thewelfare and economic e¤ects of a bilateral FTA between the United States andSouth African Customs Union (SACU). Similarly, Jordaan (2011) uses paneldata estimation, the gravity model, to determine the e¤ects of the EU-SA andSADC trade agreement.

As CGE methodology takes the intricate interaction between and within sec-tors into account, other studies analyse the impact of trade liberalisation usingvarious CGE models and databases (Gilbert and Scollay (2000) and O’Ryan etal. (2011)). To analyse the economic e¤ects of four East Asian FTA options,Kitwiwattanachai, Nelson and Reed (2010) consider unemployment, to inves-tigate the changes in real wage and subsequently unemployment due to tradeliberalisation. The FTA contributes to higher economic welfare gains comparedto any of the other bilateral trade agreements (Kitwiwattanachai, Nelson &Reed, 2010).

Several studies conduct research into the proposed TFTA by means of CGEmethodology. In a research report, Minor and Mureverwi (2013) evaluate the

6Some trade agreements exist due to political motivation and not necessarily because ofeconomic bene…ts that might arise for the participating country (Baier, Bergstrand & Clance,2014).

5

e¤ects of African trade liberalisation (the TFTA and …nally, the customs union)on Mozambique, focussing on the vulnerability of poor households. The au-thors employ the MyGTAP model and database, developed by Peter Minor andTerrie Walmsley, which allows for the deconstruction of GTAP regional house-holds. The basis of the MyGTAP model is the standard, static GTAP model.Increased trade contribute to higher real income for poor rural households, how-ever, the same is not true for poor urban households. Government revenues areexpected to increase and could be applied to compensate for the loss of poorurban households (Minor & Mureverwi, 2013).

Similarly, Mukwaya and Mold (2014) analyse the e¤ect of TFTA on theEast African region using the GTAP model and database, to analyse the statice¤ects of the proposed TFTA. The results indicate a net welfare gain of $10.7billion. The distribution of gains mainly goes to Egypt, South Africa and Zim-babwe. Increased exports and imports are expected, due to ampli…ed industrialproduction across the East African region (as new …rms enter the market space).

Using the standard comparative, static pre-release version 8 GTAP model,Jensen and Sandrey (2011) simulate a 2 percent reduction in assumed non-tari¤ barriers to both merchandise goods and services barriers. Results indicatethat South Africa and Mozambique gain most favourably from the TFTA. SouthAfrica’s exchange rate appreciates when the demand for exports increase. Indus-tries in South Africa that may experience improvement include the agriculturalsugar sector, manufacturing and services output (Jensen & Sandrey, 2011).

Likewise, using an ex-ante general equilibrium approach, Willenbockel (2013)uses a GLOBE model and considers eight simulation scenarios. GLOBE is amulti-country CGE model, developed by McDonald, Thierfelder and Robinson(2007) for the investigation of trade negotiations. Commensurate to Jensen andSandrey (2011), Willenbockel …nds that South Africa enjoys the largest incomegains from the TFTA under full Intra-Free Trade Area tari¤ liberalisation. Inthis study Swaziland, Lesotho and Namibia are expected to gain the largestrelative to the baseline (Willenbockel, 2013).

As determined by other studies, South Africa is expected to gain consider-ably from the TFTA relative to other participating countries. The aim of thisstudy is to conduct a general equilibrium analysis of the e¤ects of the TFTAon South Africa and the di¤erent participants within the economy. In orderfor policy makers to e¤ectively implement the agreement and manage subse-quent challenges that may arise, it is of essence to determine the players in theeconomy who will either bene…t or lose from such an agreement.

4 Methodology

In this study, we use a CGE model to analyse the economic e¤ects of the TFTA.Four basic tasks distinguish CGE based analysis: The …rst task is the deriva-tion of the model’s theoretical structure. The second task is calibration, whichintegrates the construction of the database, as well as the evaluation of coe¢-cients and parameters for the base year. The third task is the simulation design,

6

which includes the choice of an appropriate model closure for the given exoge-nous shock. The fourth and …nal task is the interpretation of the simulationresults, considering only the underlying database values, theory and assump-tions incorporated in the simulation design (Adams, 2005).

This method of modelling makes use of real world data, taking into accountthe inter-linkages between di¤erent sectors and agents whilst comparing thee¤ects of a certain change, relative to the baseline. CGE models simultane-ously consider all the sectors of the economy and take account of economy-widespillover e¤ects. To solve for a set of prices and quantities where demand andsupply are in equilibrium, consumer and producer behaviour with market clear-ing constraints are imposed (Bur…sher, 2011).

A CGE model is a system of equations, which describes the economy andthe interaction amongst the di¤erent sectors and participants. The foundationof a CGE model is directly derived from economic theory and includes exoge-nous and endogenous variables, in conjunction with market clearing constraints.Equations are solved simultaneously to obtain equilibrium of the economy. Sta-tic CGE models provide a before and after comparison of an economic shockor policy implementation (Bur…sher, 2011).7 The given CGE database refers toa given time period (the base, business-as-usual, reference of benchmark year),which through simulation, will then take the database from the initial or currenttime-period to the next (Ianchovichina & McDougall, 2000).

A wide variety of CGE models exist for di¤erent applications. The GTAPmodel, employed in this study, is a static multi-country CGE model that is ableto assess economic shocks and policy changes in a global trade framework. Inaddition to the GTAP model, we utilise the GTAP version 8.1 database. Abene…t of using a multi-country CGE model is the endogenous considerationof the foreign market, which recognises that policies implemented in certaincountries a¤ect other countries as in the real world.

This section further considers the tasks that distinguish CGE based analysis.Section 4.1 considers the …rst task and discusses the GTAP model; Section4.2 evaluates the GTAP database. Hereafter Section 4.3 addresses the thirdtask, which is the simulation design and model closure. Section 6 provides theinterpretation of the results of the simulation, which is the …nal task.

4.1 Global Trade Analysis Project (GTAP) Model Overview

The GTAP was established in 1992 to develop a global model and databasein order to quantitatively analyse international economics within an economy-wide framework. The GTAP model is a comparative-static multi-country ap-plied model of the world economy to use in conjunction with the multi-countrydatabase – the GTAP database. This multi-country model was created in or-der to analyse bilateral and multilateral trade agreements, immigration, climatechange and international …nancial imbalances (Hertel, 2013).

7Conversely, dynamic CGE models are solved recursively, to take into account the adjust-ment of variables and factors at each time-period and provides results that are more realistic.

7

The design of the GTAP model entails the symmetric handling of utility andproduction functions across di¤erent regions. In each region, the representativehousehold maximises a Cobb-Douglas utility function by allocating income be-tween private and public consumption, as well as savings.8 Households investtheir savings locally or abroad in the global trust, also referred to as the globalbank. Private consumption behaviour is modelled using a constant di¤erenceof elasticity (CDE) function, where the demand elasticity di¤ers in response tochanges in expenditure and prices. Correspondingly, government expenditureis also represented by a Cobb-Douglas utility function. Although utility andproduction functions are treated symmetrically, di¤erences in regional behav-iour are captured in the di¤erences of economic ‡ows, model parameters andthe model closure.

Figure 4 is an illustration of the economic relationships in the model. Re-gional households or consumers earn income through rent of capital, labour, landand natural resources (primary factors) that they own (VOA endowments) andprovide to producers. Another form of remuneration includes taxes (TAXES)and tari¤s on imports, as well as exports (MTAX and XTAX respectively). Theregional household allocates expenditure between private expenditure (PRIV-EXP), government expenditure (GOVEXP) and savings (SAVE). This includesconsumption of domestic goods (VDPA and VDGA) and imported goods (VIPAand BEAM), which is included in import tari¤s (MTAX), consumer tax (TAXES)and savings (SAVE).9

Producers provide goods and services to international and domestic economicparticipants and obtain revenue from these di¤erent sources. Revenue for theproducer constitutes out of the total value of government and private sector atmarket prices (VDGA and VDPA respectively), as well as intermediate con-sumption between producers (VDFA) and exports to international participants(VXMD). Producers buy inputs and primary resources from private households(VDPA), which is included in taxes paid (TAXES) and imports (VIFA) (Hertel& Tsigas, 1997).

Pro…t–maximising producers base output decisions on a constant elasticityof substitution (CES) function, represented in the nested production structureillustrated in Figure 5. At each level of production, the demand for commoditiesdepends on the elasticity of substitution and relative prices of inputs. Commodi-ties are produced with CES between primary factors and intermediate inputs.Intermediate inputs may originate from domestic or foreign resources, based onthe Armington assumption.

We use the GTAP version 3.6 model to study the impact of the TFTA on theSouth African economy. Together with the GTAP model and database, we usethe GEMPACK suite of economic-modelling software that includes RunGTAP,developed by the Centre of Policy Studies (Pearson & Horridge, 2003). InSection 4.2, we present the GTAP database, which involves the third task ofCGE analysis.

8A special case of Stone-Geary utility function is employed, due to all subsistence sharesequaling zero, the function reduces to the Cobb-Douglas function.

9Households save through the net investment (NETINV) of their capital in producers.

8

4.2 Global Trade Analysis Database

A CGE database portrays the circular ‡ow of income and expenditure in aneconomy at a speci…c time-period. A CGE database reports the aggregated val-ues of all goods and services that the economy produces and the correspondingincome generated. We employ version 8.1 of the GTAP database with 2007 asthe reference year.10

The GTAP database records the annual ‡ows of goods and services for theglobal economy in the benchmark year. With linkages between industries of eachregional economy already in place in the database, each economy is consolidatedwith private and government consumption, investment, exports and imports(James & McDougall, 1993). To model the global economy accurately, theGTAP database considers and incorporates bilateral trade preferences, tari¤rate quotas (ad valorem and speci…c tari¤s individually), export subsidies andagricultural support (Hertel, 2013).

The full GTAP database contains 134 regions for all 57 GTAP commoditiesand is documented in Aguiar, Narayanan and Walmsley (2012). For this study,we aggregate the 134 regions into seven regions: The Republic of South Africa(ZAF), the rest of countries participating in the TFTA (Other TFTA), the rest ofAfrica (RoA), the North American Free Trade Area (NAFTA), the EU (EU25),the BRIC countries (BRIC) and the Rest of the World (RoW).11 We aggregatethe 57 commodities (in‡ow and out‡ow) into fourteen sectors (tabulated inTable 3) in order to investigate the micro economic impact of the TFTA.12

To produce a consistent economic model, the GTAP database has to adhereto certain requirements. Exports of a certain economy are imported by othereconomies. Regional economies function on their respective budget constraints.Sectors earn zero pro…t. Global savings should equal global net investment.Furthermore, exports of global transport services from an individual countrymust be equal to the demand for the same services (Hertel, 2013). In Section4.3 we discuss the model closure and simulation design.

4.3 Model Closure and Simulation Design

Similar to other CGE models, the GTAP model contains more variables thanequations. In order to solve the model, it is necessary to specify variables tobe determined endogenously within the model and variables to be determinedexogenously The model closure is therefore the assumptions regarding the choiceand speci…cation of endogenous and exogenous variables. The model closureportrays the desired economic environment in which the simulation is to be run(Dixon et al., 2013).

The standard GTAP closure is a short-run closure. The economic environ-ment is characterised by:

10The GTAP version 8.1 database was developed; maintained and is available from PurdueUniversity, USA.

11The detailed aggregation of regions is available in Appendix A, Table A1.12The detailed aggregation of commodities is available in Appendix A, Table A2.

9

² Level of activity in endowment sectors is exogenously determined.13

² Aggregate employment of labour (skilled and unskilled) and capital is…xed.

² Government budget balances are slack.14

² Sluggish endowment commodities such as land and natural resources aredynamically de…ned.

² Investment and capital in each region move together.

² Global rate of return adjust to ensure each region’s investment equals thechange in global savings.

² All factor technical e¢ciency (technological change) is exogenous and heldconstant indicating that the policy shock will not cause evolution of tech-nology and productivity15 (Hertel & Tsigas, 1997).

The GTAP policy simulation is the removal of all ad valorem import tari¤sbetween South Africa and other TFTA countries. Tari¤s and other import pro-tection between the TFTA, South Africa and other regions remain unchanged.

The shock therefore has two separate elements: Firstly, the elimination oftari¤s levied on South African goods imported into the TFTA. Secondly, theelimination of tari¤s levied on TFTA goods imported into South Africa.

When analysing the impact of the TFTA on the South African economy,it is important to keep the model closure and simulation design in mind. Thefollowing section presents a small back-of-the-envelope model to assist in theinterpretation of results, which follows in Section 6.

5 Back-of-the-Envelope Model

A simple back-of-the-envelope (BOTE) model assists with the interpretationand understanding of simulation results. This stylised model of economic rela-tionships supports the evaluation of economic e¤ects that occur because of theelimination of tari¤s between South Africa and other TFTA countries.

The BOTE model identi…es the primary theoretical mechanisms that moti-vate the projection and results of the full model. The model also emphasisescentral elements of the database (Adams, 2005). The stylised model developed

13Capital stock is assumed to be constant as capital is immobile in the short-run. The rateof return is therefore determined endogenously in order for the capital market to clear. Thereal rate of return of all primary factors is determined endogenously.

14The GTAP model does not link government expenditure to government revenue. Govern-ment expenditure depends on regional income, which is allocated between private consumptionexpenditure, public consumption expenditure and savings.

15 If it is believed that the policy shock will a¤ect technology evolution and the productivityof a country, an additional shock can be implemented to take this into account.

10

by Adams (2005) has a multi-country dimension, which is simpli…ed in this pa-per to evaluate only the main economic variables of interest. The BOTE modelis depicted in Table 4.

Equation (1) de…nes the real GDP of region r (GDP (r)), which is equalto the sum of real private consumption (C(r)), real investment (I(r)), realgovernment consumption (G(r)) and the net volume of exports (which is thevolume of exports (X(r)) less the volume of imports (M(r))). The volume ofimports is determined by real GDP, the terms of trade (TOT (r)) and the averagepower of the average rate of tari¤s (T ) in equation (2). Equation (3) depictsthe volume of exports, which is determined by the general level of economicactivity (GDP ) and inversely a¤ected by the terms of trade. South Africafaces a downward sloping curve for export-demand due to the assumption ofArmington elasticities. In this analysis, the increase in the export demand forSouth African goods will shift this downward sloping curve to the right.

Terms of trade is equal to the price of exports (P (r)x) relative to the price ofimports (P (r)M). This, we show in equation (4). The rate of return (ROR(r))of the various endowments in the economy is depicted in equation (5) as thefunction of the capital to labour ratio in the region (K

L(r)), the terms of trade

in that region and the exogenous level of technology (A), which is assumedexogenous. The various economic relationships discussed in this section aid toelucidate how the shock enters the market, ripples through the economy andcreate numerous economic consequences.

The following section presents the macroeconomic results, South Africanspeci…c macroeconomic results and industry results.

6 Simulation Results

This section presents the …nal task of CGE analysis - the interpretation ofsimulation results. Simulation results represent percentage deviations from thebaseline. For instance, if it were forecasted that South Africa’s real GDP willincrease by 5 percent in the coming year, from its current level of R3.5 trillionto R3.675 trillion, in a business-as-usual scenario16 , this would be consideredas the baseline. If the model predicts that the impact of the exogenous policyshock would increase real GDP growth in the short run by 1 percent, this wouldbe equivalent to predicting that real GDP increase to R3.712 trillion instead ofR3.675 trillion, because of the policy change.

With technological change held constant in the closure, the percentage changein technical e¢ciency variables are zero. That is, we do not allow for the imple-mentation of the TFTA in the policy simulation to change the path productivitygrowth, relative to the baseline. The same is true for other variables held con-stant in the closure. In order to incorporate the possible changes that mighttake place as a because of say, technical e¢ciency improving as a result of the

16This is the baseline situation and does not include the exogenous shock under considera-tion.

11

policy change, a separate shock to the relevant exogenous variable would haveto be implemented.

This section considers the impact of the policy shock and is structured as fol-lows: Section 6.1 discusses macroeconomic and welfare results for the respectiveregions; Section 6.2 interrogates South Africa speci…c macroeconomic resultsand Section 6.3 looks at selected South African industry results.

6.1 Overview of Macroeconomic and Welfare Results

South African GDP increases due to the establishment of the TFTA, whilstother TFTA countries experience a relatively large decrease in value of GDPcompared to other regions not participating in the TFTA.17 Certain countriesparticipating in the TFTA, may enjoy increased value of GDP and welfare,however due to the aggregation of the full GTAP database, we do not evaluatethe e¤ects on individual countries.18 As expected, the impact of the TFTA onother regions not participating in the TFTA is negative, yet negligible. As tech-nical e¢ciency is assumed constant, the model does not allow for technologicalspillovers between countries. If the model did allow for such spillovers, we mayobserve an increase in the value of GDP in other TFTA countries and regionsnot participating in the trade agreement.

The increase of South Africa’s value of GDP by 1.023 percent may be dueto the change in patterns of imports and exports, illustrated in Appendix B.19

The volume of merchandise exports in South Africa and other TFTA coun-tries increase signi…cantly by 0.685 percent and 0.547 percent, respectively. Thevolume of merchandise imports increase in South Africa and other TFTA coun-tries, however decrease in all other regions not participating in the TFTA. AsSouth Africa’s terms of trade improve by 0.698 percent, imports become rela-tively less expensive, whereas exports become relatively more expensive. Thenagain, other TFTA countries can now import at lower cost from South Africa,contributing to the overall increase in the region’s imports.20

Output of the capital goods sector increases in South Africa and other TFTAcountries by 1.155 percent and 0.357 percent, respectively. We can attribute

17Countries participate in trade agreements due to various reasons such as economic and/orpolitical bene…ts, or due to pressures from other countries to take part. All countries partici-pating in a trade agreement, will therefore not necessarily gain in terms of GDP and welfare.

18The aim of the paper is to investigate the impact of the TFTA on the South Africaneconomy, therefore for simplicity all other participants in the TFTA were aggregated into onesingle region as discussed in Section 4.2 and Appendix A.

19Change in volume of total exports from region r to destination s is provided in AppendixB, Table B1. The volume of total exports from South Africa to all regions except other TFTAcountries decrease relative to the baseline, indicating that South Africa diverts trade away fromother regions. Other TFTA countries export more to global trading partners, however less toother countries participating in the TFTA. On the other hand, the rest of Africa decreasesits exports to all regions, except South Africa and other African countries not included in theTFTA. South Africa also resorts in importing more from other African countries due to theterms of trade improvement.

20Other TFTA countries now substitute goods imported from BRIC countries with importsfrom South Africa, explaining the decrease in the exports of the BRIC region 0.002 percent.

12

this to the increase in production and GDP in certain secondary industries.The market price for primary factors increase in South Africa by 1.004 percent,whilst other regions experience a decrease. As the terms of trade improve inSouth Africa and tari¤s decrease, the real price of labour, capital and otherendowments increase, escalating the market price index for all primary factors.21

We examine this further in Section 6.2 and Section 6.3.Furthermore, the establishment of the TFTA a¤ects welfare in the di¤erent

regions speci…ed in the model. The decomposition of welfare e¤ects allows forthe identi…cation of the welfare contributions by commodity, factor and taxtype (Hu¤ & Hertel, 2000). Global welfare increases with US$ 0.805 million, asillustrated in Table 6. South Africa is the only region that experience welfaregains, whereas other TFTA countries bear a net welfare loss. The welfare gainof US$ 815.985 million for South Africa is primarily driven by the terms oftrade e¤ect. The terms of trade e¤ect contributes to decreased welfare for allregions, expect for South Africa, where the terms of trade e¤ect contributes US$602.899 million.22 The total investment-savings e¤ect contributes US$ 19.930million to increased welfare. Furthermore, South Africa experiences a welfaregain through the allocative e¢ciency e¤ect, whilst this e¤ect decreases for otherTFTA countries, as a region, relative to the baseline.

The BRIC countries and the rest of the world also lose, relative to thebaseline, in this regard. The net welfare losses in rest of Africa and BRIC aredriven by the loss of the total investment-savings e¤ect. Although the totalterms of trade e¤ect is globally negative, the total loss of global welfare isprimarily driven by the negative impact of the trade tax e¤ect of US$ 13.760million and the total investment-savings e¤ect of US$ 0.060 million.

Considering the welfare decomposition, the TFTA has a negative impact onthe welfare of other countries participating in the TFTA as a single region. Inaddition, the rest of Africa not included in the TFTA loses in terms of welfare.The TFTA would thus prove to be more bene…cial for South Africa than toother TFTA countries (as a region) participating. Section 6.2 investigates andinterprets South African macroeconomic results.

6.2 South African Macroeconomic Results

To reiterate, the model is shocked with the elimination of import tari¤s betweenSouth Africa and other TFTA countries. The trade pro…le presented in Section2.2 and the BOTE model discussed in Section 5 assist in the interpretation ofresults, presented in Table 7. The …rst round impact of the elimination of import

21The positive shift in the demand for South African goods causes increased pressures onendowments used in production. Due to the assumption of exogenous technical e¢ciency,prices of endowments increase to show the increased demand and supply.

22Again, note that some countries participating in the TFTA can gain, however the overalle¤ect is observed in this paper. Some countries may gain signi…cantly, however the lossexperienced by others outweighs the bene…ts for others. In the paper by Jensen and Sandrey(2011), the authors …nd that only South Africa and Mozambique gain in terms of welfare.Mukwaya and Mold (2014) …nd that net welfare gains procure to Egypt, South Africa andZimbabwe.

13

tari¤s is expected to increase the demand for imports (in South Africa andother TFTA countries) at the expense of demand for domestic products. As theagreement eliminates tari¤s between South Africa and other TFTA countries,trade increases and positively a¤ect the South African economy.

The selected macroeconomic variables presented in Table 7 show an im-provement after the shock of the elimination of tari¤s is applied. South Africa’seconomy grows at a relative higher rate, as a result of the implementation ofthe TFTA. The GDP quantity index increases by 0.068 percent and the value ofGDP increases by 1.023 percent relative to the baseline. Moreover, the increasedGDP23 contributes to higher levels of private consumption (1.082 percent) andgovernment consumption expenditure (1.120 percent).24

The elimination of tari¤s between South Africa and other TFTA countriesdirectly decreases the price of imports for South Africa by 0.012 percent. Con-comitantly, the price of imports from South Africa into other TFTA countriesdecrease, leading to increased demand for South African exports. The positiveshift in the downward sloping demand for South African exports (more so thanfor other TFTA countries as a region), bring about an increase the price ofexports of 0.686 percent and contribute to improved terms of trade.25 With anincrease in the price of exports and a decrease in the price of imports, SouthAfrica’s terms of trade improve by 0.689 percent.26 Increased GDP, improvedterms of trade and reduced ad valorem import tari¤s contribute to the increasein South African imports, as seen theoretically in equation (2). From the simula-tion results, we observe the increase in volume of exports and imports increasingby 0.685 percent and 2.003 percent, respectively. The relatively higher increasein imports than exports contributes to the deterioration of the trade balance byUS$ 596.769 million.

The increase of South Africa’s terms of trade positively a¤ects the rate ofreturn to endowments, contributing to an increase in the prices of endowments.Equation (5) depicts the rate of return as a function of the ratio of capital andlabour, technology (both of which are assumed exogenous in the model closure)and the terms of trade. The current rate of return on capital stock increase of0.685 percent is thus due to the terms of trade improvement. The rate of returnto capital contributes to the increase in the real price of capital of 0.291 percent.The market price index for primary factors increase by 1.004 percent, relativeto the baseline. Moreover, the real price of unskilled and skilled labour increase

23 Increased GDP signi…es increased regional expenditure, which is equal to increased re-gional income.

24Government budget balance deteriorates due to the loss of tari¤ revenue. This does notimply that government expenditure decreases or increases in other taxes. As reductions intaxes and tari¤s lead to reduced excess burden, regional real income increases contributeto increased government (public consumption) expenditure (Hertel & Tsigas, 1997). Mostpublished GTAP applications assume this (Adams, 2005).

25From the simulation results, it is observed that South Africa shifts exports away fromother regions not included in the TFTA, toward other countries participating in the TFTA.The overall supply of exports to all regions thus decreases. From basic international tradebackground, it is known that with decreased supply of South African exports (the shift of thesupply curve to the left) – the price of exports also increases.

26Refer to equation (4) in the BOTE model.

14

by 0.314 and 0.338 percent, respectively.27

With the real price for labour (unskilled and skilled) increasing relativelymore than capital, it is expected that output of capital goods sector rise, due toincreased demand for capital goods. An increase in output in the capital goodssector of 1.155 percent represents an increase in investment.

Overall, the improvement in trade, investment and expenditure contributesto the increase of GDP viewed from the expenditure side, as in equation (1).Although the overall South African economy bene…ts from the implementationof the TFTA, certain industries may lose as a result of macroeconomic changes.Section 6.3 presents the various impacts of the TFTA on the di¤erent industriesof South Africa. This section also investigates the demand for the respectiveendowments in various industries.

6.3 Industry Results

Due to the increase in imports in all industries, industries substitute domesti-cally produced goods with imports. This is true for industries with negligibleexports to other TFTA countries. Output (value added) in most industriesdecrease, however, secondary industries such as processed food, textiles andclothing, petroleum and coal products, as well as construction experience anincrease in output. From Section 2.2, we observe that processed food as well aspetroleum and coal products contribute 20 percent of exports to other TFTAcountries. As processed foods and petroleum and coal products were highlytaxed by other TFTA countries - the elimination of tari¤s motivates the pro-duction and consequently, the export of these goods to other TFTA countries.Output, imports, exports and selected price indices for respective industries aretabulated in Table 8.

The percentage change in private and government demand for domestic andimported goods may provide us with insight into output changes within indus-tries.28 Private and government demand for imported primary industry goodssuch as grain and crops, livestock and meat products, as well as secondary in-dustry goods such as textile and clothing increase most notably compared toother goods.

Conversely, the domestic demand for grains and crops, textiles and cloth-ing as well as manufactured goods decrease relative to the baseline. This isindicative of the substitution of domestic goods with imported goods. Thissubstitution of domestic goods with imported goods, explains the decrease inoutput of particular industries.

In all industries the terms of trade e¤ect is observable as the increase inprivate and government demand for imported goods, exceed that of domesti-cally produced goods. Due to the decreased price of imports, it becomes morea¤ordable to import certain goods than to produce domestically. This is thecase for all goods, except for coal extraction, where an increase of 0.021 percent

27This is measured as the ratio of return of endowment to in‡ation (CPI).28See Figure 6.

15

occurs in the price of imports. As a result, South Africa experiences an increasein overall imports in the all industries. Demand for goods imported mainlyfrom other TFTA countries into South Africa such as oil and gas extraction andmanufactured goods, increases.29

The ratio of domestic to import prices is positive in all industries, implyingthat domestically produced goods are more expensive than imported goods.30

The ratio of domestic to import prices for grains and crops increases signi…cantlyby 1.182 percent, relative to that of other goods. Therefore, it is economicallysensible for industries to substitute locally produced goods with less expensiveimported grains and crops. As South Africa’s terms of trade improve and thecurrency appreciates, imports become cheaper and replace expensive domesti-cally produced goods.

Imports and exports determine domestic production in industries. The elimi-nation of tari¤s most favourably a¤ects high exporting industries and in speci…c,industries where exports to other TFTA countries are high. Secondary indus-try goods such as processed food, petroleum and coal products, textiles andclothing and, to a lesser extent, manufactured (light and heavy) goods exportsincrease respectively by 9.771 percent, 10.272 percent, 7.940 percent and 0.694percent.31 As other TFTA countries highly taxed processed foods imports fromSouth Africa, it sheds light on the sharp increase in exports to other TFTAcountries.32

Exports to other TFTA countries increase considerably, yet at the sametime decrease to other trading partners, relative to the baseline, as depicted inFigure 7. Oil and gas extraction exports to other TFTA countries increase by35.200 percent, whilst exports to other countries such as North America andthe European Union decrease by 3.433 and 3.599 percent. Other industries thatexperience increase in exports sales to other TFTA countries are processed food(31.863 percent), textiles and clothing (32.030 percent), as well as light andmanufactured goods (27.466 percent).33

The TFTA as a trade agreement has a noteworthy impact on trade in allindustries, a¤ecting the level of production. The changes in production levelscause ‡uctuations in the demand for endowments in industries. In order tobene…t from potential improvements or address possible challenges, we considerthe changes in demand for endowments. Table 9 reports the percentage changein the demand for the …ve primary factors by the di¤erent industries. In thismodel, we consider the activity (expansion or contraction) of endowment sectorsas exogenous.

29See Figure 2.30The ratio of domestic to import prices compares the price of domestically produced goods

to that of imported goods.31Although a substantial increase in exports of petroleum and coal products is experienced,

total exports of this product are small and equal only 4.235 percent of all exports (GTAPDatabase).

32 37.055 percent of processed foods are exported to other TFTA countries and 26.579 percentto the European Union.

33Manufactured goods make up 69.742 percent of all exports to the TFTA, while processedfood and textiles and clothing also being of the high exported goods to this region.

16

The overall demand for endowments by industries increases, relative to thebaseline. Total South African demand for endowments increases by 10.104 per-cent.34 Demand for unskilled and skilled labour increase by 2.842 and 2.502percent, respectively.35 The grains and crops industry demands less labour,land and capital endowments, because of the decreased output resulting fromimport substitution. With production in industries contracting, we do not ex-pect an expansion in the demand for endowments that these industries cannotoptimally employ.

The demand for labour in primary industries concerned with mining activ-ities, as well as electricity and gas, decrease considerably. As skilled labour iscomparatively more expensive than unskilled labour, the decrease in demand forskilled labour decreases by a higher margin. Demand for employment by extrac-tion industries decrease signi…cantly relative to other industries. Demand forunskilled and skilled labour by the other extraction industry decrease by 0.785percent and 0.789 percent, in the coal extraction industry by 0.586 percent and0.591 percent and in the oil and gas extraction industry by 0.517 percent and0.521 percent, respectively. Conversely, industries such as processed food, pe-troleum and coal products, water, construction and other services demand moreunskilled and skilled labour. Industries that experience an increase in demandfor labour, also experience increase in demand for all other primary factors.

Furthermore, the increases in the use and demand of capital denote theinvestment expenditure in industries and the economy. Demand for capitalincreases considerably in the petroleum and coal (1.995 percent), processed food(1.226 percent) and construction (0.993 percent) industries. The increase indemand for capital in these secondary industries signi…es the expansion andgrowth because of the TFTA and explains the overall increase in the demandfor capital of 3.172 percent, relative to the baseline. In a growing economy, weexpect the increase in the demand for all primary factors, as well as improvementand increased investments in infrastructure and accompanying services.

From the analysis of the results, it is apparent that secondary industriesbene…t from the TFTA, whilst primary industries lose. Although overall demandfor labour increases; structural unemployment may possibly arise as the demandin agricultural and mining industries decrease. Industry speci…c economic e¤ectsare of importance to consider when developing policies. Section 7 providesconcluding remarks and discusses possible considerations when evaluating thee¤ects of the TFTA on the South African economy.

34This result may seem high relative to other macroeconomic results however; the percentagechange represents the aggregated demand for endowments of all industries. The percentagechange in the total demand for all endowments is thus calculated by considering the changein demand for a certain endowment relative to the share of that endowment employed in theindustry.

35The percentage deviation in the demand for natural resources is positive overall, yetrelatively small.

17

7 Conclusion and Policy ImplicationsWith the South African National Planning Commission emphasising the im-portance of increased trade between South Africa and other African countries,the e¤ects of the TFTA is of great signi…cance to policymakers. In this paper,we use a global CGE model to analyse the impact of the TFTA on the SouthAfrican economy.

Simulation results indicate that South Africa gains relative to other TFTAcountries, as a region. South Africa’s value of GDP increases by 1.023 percent,relative to the baseline. The boom in South African exports contributes to theimprovement of the terms of trade of 0.698 percent. As South Africa increaseexports to other TFTA countries, exports to other non-participating regions de-crease. Although trade diversion is expected, South Africa should guard againstthe neglect of existing trade relationships with major trading partners. Withoverall imports and exports increasing because of the increase in demand, SouthAfrica’s industries are a¤ected di¤erently. The TFTA brings about increasedoutput in secondary industries such as processed food, petroleum and coal prod-ucts.

Increased outputs in these industries consequently contribute to increaseddemand for endowments, such as skilled and unskilled labour, capital and land.Other industries experience substitution of domestically produced goods withthat of now relatively less expensive imported goods. Private and governmentdemand for grains and crops, livestock and meat products, as well as textile andclothing industries increase notably with the demand for domestically producedgoods decreasing relative to the baseline.

In contrast to the boom created on the export side, the substitution of do-mestically produced goods with imported goods by some users slightly reducesoutput again, and subsequently, the demand for endowments. Although over-all demand for employment increases by 2.842 percent for unskilled and 2.502percent for skilled labour, primary industries such as coal extraction, oil andgas extraction and other extractions demand less labour, which may give riseto structural unemployment for these type of workers.

In order to take advantage of the bene…ts and minimise negative repercus-sions, government should adapt economic policies accordingly. To compensatefor possible increased structural unemployment, policymakers can react by re-ducing restrictive labour laws that will encourage employment of both skilledand unskilled labour in the particular industries where demand for labour de-creases. As unemployment is a macroeconomic concern for South Africa, suchan intervention is well worth investigating for all industries. To motivate eco-nomic activity in industries where production decreases, government could easethe tax burden and red tape on companies operating in these industries, whichmay result in increased production activity.

The overall impact of the policy change on the South African economy ispositive and South Africa can draw obvious bene…ts from the establishmentof this agreement. We expect the TFTA to increase economic growth, trade,

18

consumption and government expenditure36, investment and employment inSouth Africa. The economic e¤ects of the TFTA on South Africa add to theNDP vision to increase global participation and economic cooperation, in ane¤ort to advance the South African economy for all.

References

[1] ABEDINI, J. and PERIDY, N. (2014). Trade E¤ects of Regional Integra-tion in Imperfect Competition: Evidence from the Greater Arab Free TradeArea (GAFTA). International Economic Journal, 28(2): 273-292.

[2] ADAMS, P. D. (2005). Interpretation of results from CGE models such asGTAP. Journal of Policy Modelling, 27(8): 941-959.

[3] AGUIAR, A. H., NARAYANAN, B. and WALMSLEY, T. L. (2012). In-troduction to the Global Trade Analysis Project and the GTAP Data Base.GTAP Working Paper No. 3965, Purdue University, Center for GlobalTrade Analysis.

[4] BAIER, S. L. and BERGSTRAND, J. H. (2004). Economic determinantsof free trade agreements. Journal of International Economics, 64(1): 29-63.

[5] BAIER, S. L. and BERGSTRAND, J. H. (2007). Do free trade agreementsactually increase members’ international trade?. Journal of InternationalEconomics, 71(1): 72-95.

[6] BAIER, S. L., BERGSTRAND, J. H. and CLANCE, M. W. (2014). Pre-liminary examination of heterogenenous e¤ects on international trade ofeconomic integration agreements. In Dur, A. and Elsig, M. (ed.), TradeCooperation: The Purpose, Design and E¤ects of Preferential Trade Agree-ments. Cambridge: Cambridge University Press, 355 – 373.

[7] BAIER, S. L., BERGSTRAND, J. H. and FENG, M. (2014). Economicintegration agreements and the margins of international trade. Journal ofInternational Economics, 93(2): 339-350.

[8] BEN-DAVID, D. and LOEWY, M. B. (1998). Free Trade, Growth, andCovergence. Journal of Economic Growth, 3(2): 143-170.

[9] BOHLMANN, H. R., DIXON, P. B., RIMMER, M. T. and VAN HEER-DEN, J. H. (2014). The Impact of the 2014 Platinum Mining Strike inSouth Africa: An Economy-Wide Analysis. Working Paper 478, ERSA.

[10] BROWN, D. K., KIYOTA, K. and STERN, R. M. (2008). An Analysisof a US-Southern African Customs Union (SACU) Free Trade Agreement.World Development, 36(3): 461-484.

36As a result of increased economic growth and activity.

19

[11] BURFISHER, M.E. (2011). Introduction to Computable General Equilib-rium Models. Maryland: Cambridge University Press.

[12] BUSTOS, P. (2011). Trade Liberalization, Exports and Technology Up-grading: Evidence on the Impact of MERCOSUR on Argentinian Firms.American Economic Review, 101(1): 304-340.

[13] CERNAT, L. (2003). Assessing South-South regional integration: same is-sues, same metrics. Issue 21, Policy Issues in International Trade and Com-modities Study Series.

[14] COMESA-EAC-SADC TRIPARTITE. (2014). COMESA-EAC-SADC Tri-partite. [Online] Available at: http://www.comesa-eac-sadc-tripartite.org/[Accessed 10 November 2014].

[15] ENSOR, L. (2014). SA bene…ts from the SADC deal with EU. [Online]Available at: http://www.bdlive.co.za/business/trade/2014/07/02/sa-bene…ts-from-sadc-deal-with-eu[Accessed 16 August 2014].

[16] GILBERT, J. and SCOLLAY, R. (2000). Measuring the Gains from APECTrade Liberalisation: An Overview of CGE Assessments. The World Econ-omy, 23(2): 175-197.

[17] HERTEL, T. W. (2013). Global Applied General Equilibrium Analysis Us-ing the Global Trade Analysis Project Framework. In Dixon P.B. and Jor-genson, D.W. (ed.), Handbook of Computable General Equilibrium Model-ing, Volume 1. United States:North Holland, 815-876.

[18] HERTEL, T. W., HUMMELS, D., IVANIC, M. and KEENEY, R. (2007).How con…dent can we be of CGE-based assessments of Free Trade Agree-ments. Economic Modelling, 24(4): 661-635.

[19] HERTEL, T. W. and TSIGAS, M. E. (1997). Structure of the GTAP. InHertel, T.W. (ed.), Global Trade Analysis: Modeling and Applications. Lon-don: Cambridge University Press: London, 13 - 73.

[20] HUFF, K. M. and HERTEL, T. W. (2000). Decomposing Welfare Changesin the GTAP Model. GTAP Technical Paper No.5, Purdue University, Cen-ter for Global Trade Analysis.

[21] IANCHOVICHINA, E. and HERTEL, T. M. R. (1999). The East AsianEconomic Crisis: It’s Not All Bad News. Choices, Second Quarter, 15 – 22.

[22] IANCHOVICHINA, E. and MCDOUGALL, R. (2000). Theoretical Struc-ture of Dynamic GTAP. GTAP Technical Paper No. 17, Purdue University,Center for Global Trade Analysis.

[23] JAMES, M. and MCDOUGALL, R. (1993). An input-output data updatefacility for SALTER. SALTER Working Paper 17, Industry Commission,Canberra.

20

[24] JENSEN, H. G. and SANDREY, R. (2011). The tripartite Free TradeAgreement: A computer analysis of the impacts, tralac Working Paper No.IIWP06/2011, Stellenbosch: tralac.

[25] JORDAAN, A. (2011). Analysing the Trade E¤ects of the EU-SA & SADCTrading Agreements: A Panel Data Approach. South African Journal ofEconomic and Management Sciences, 14(2): 229-244.

[26] KING, A. (2012). Economy-Wide Impacts of Industry Policy. Working Pa-per No. 12/05, New Zealand Treasury.

[27] KISHTAINY, N. (ed.) (2012). The Economics Book. London: DorlingKindersley Limited.

[28] KITWIWATTANACHAI, A., NELSON, D. and REED, G. (2010). Quan-titative impacts of alternative East Asia Free Trade Areas: A ComputableGeneral Equilibrium (CGE) assessment. Journal of Policy Modeling, 32(2):286-301.

[29] KUMAR, S. K. (2012). A GTAP Analysis of the Proposed BRICS FreeTrade Agreement: With Special Reference to the Indian Economy. 2012Conference paper, 15th Annual Conference on Global Economic Analysis,Geneva, Switzerland.

[30] MCDONALD, S., THIERFELDER, K. and ROBINSON, S. (2007).GLOBE: A SAM Based Global CGE Model using GTAP Data, Depart-mental Working Paper 14, US Naval Academy Department of Economics.

[31] MCDONALD, S. and WALMSLEY, T. L. (2003). Bilateral Free TradeAgreements and Customs Unions: The Impact of the EU Republic of SouthAfrica Free Trade Agreement on Botswana. GTAP Working Paper 29, Pur-due University, Center for Global Trade Analysis.

[32] MINOR, P. and MUREVERWI, B. (2013). A Household Level Analysis ofAfrican Trade Liberalization: The Case of Mozambique. Research Report,World Bank and Bank of the Netherlands Partnership Program.

[33] MUKWAYA, R. and MOLD, A. (2014). E¤ect of the COMESA-SADC-EAC FTA on the East African Region: Towards a New Economic Geogra-phy. 2014 Conference Paper, 17th Annual Conference on Global EconomicAnalysis, Dakar, Senegal.

[34] NARAYANAN, G. B., AGUIAR, A. H. and MCDOUGALL, R. A. (2012).Global Trade, Assistance, and Production: The GTAP 8 Data Base. Pur-due University: Center for Global Trade Analysis.

[35] NATIONAL PLANNING COMMISSION (NPC). (2011). Chapter 7: Po-sitioning South Africa in the World. In NPC, National Development Plan2030. Pretoria: Government Printer, 235 – 257.

21

[36] NATIONAL TREASURY. (2015). Chapter 2: Economic Outlook. In Na-tional Treasury, 2015 Budget Review. Pretoria: Government Printer, 15 –28.

[37] O’RYAN, R., DE MIGUEL, C. J., MILLER, S. and PEREIRA, M. (2011).The Socioeconomic and enviromental e¤ects of free trade agreements: a dy-namic CGE analysis for Chile. Environment and Development Economics,16(3): 305-327.

[38] PEARSON, K. and HORRIDGE, M. (2003). Hands-on Computing withRunGTAP and WinGEM to Introduce GTAP and GEMPACK. MonashUniversity: Centre of Policies Studies.

[39] SOUTH AFRICAN RESERVE BANK (SARB). (2014). Online statisticalquery (historical macroeconomic timeseries information). [Online]Available at: https://www.resbank.co.za/Publications/QuarterlyBulletins/Pages/QBOnlinestatsquery.aspx[Accessed 10 November 2014].

[40] STRUTT, A. and WALMSLEY, T. (2011). Implications of the Global Fi-nancial Crisis for China: A Dynamic CGE Analysis to 2020. ResearchArticle, Economics Research International.

[41] THE DEPARTMENT OF TRADE AND INDUSTRY (the dti). (2014a).Annual Report 2013/2014, Pretoria: the dti.

[42] THE DEPARTMENT OF TRADE AND INDUSTRY (the dti). (2014b).Trade, Exports & Investment. [Online]Available at: http://www.thedti.gov.za/trade_investment/ited_trade_agreement.jsp[Accessed 24 October 2014].

[43] WACZIARG, R. and WELCH, K. H. (2008). Trade Liberalization andGrowth: New Evidence. The World Bank Economic Review, 22(2): 187-231.

[44] WILLENBOCKEL, D. (2013). General Equilibrium Analysis of theCOMESA-EAC-SADC Tripartite FTA. Final Report, Brighton -UK:Institute of Development Studies at the University of Sussex.

[45] WORLD BANK. (2014). World Development Indicators. [Online]Available at: http://databank.worldbank.org/ddp/home.do?Step=3&id=4:World dataBank.[Accessed 10 November 2014].

22

Table 1: South Africa’s Main Trade Agreements

Agreement Participants

Free Trade Agreements (FTAs)

Southern African Development Community (SA

DC) 12 SADC member states

Trade, Development and Cooperation Agreement

(TDCA) South Africa and the European Union (EU)

EFTA-SACU European Free Trade Association (EFTA) and SA

CU

Customs Union

Southern African Customs Union South Africa, Botswana, Lesotho, Namibia and Sw

aziland

Preferential Trade Agreements (PTAs)

SACU-Southern Common Market (MERCOSUR

) Argentina, Brazil, Paraguay, Uruguay and SACU

Zimbabwe-South Africa South Africa and Zimbabwe

SACU-India India and SACU (still under negotiations)

Non-reciprocal Trade Agreements

Africa Growth and Opportunity Act Unilateral assistance measure granted by US to 39

Sub-Saharan African countries

Generalised System of Preference (GSP)

Unilateral Preferences offered to South Africa by t

he EU, Norway, Switzerland, Russia, Turkey, US,

Canada and Japan

Other Trade Agreements

Trade and Investment Framework Agreement (TI

FA) Bilateral agreement between US and South Africa

Trade, Investment and Development Cooperation

Agreement (TIDCA) Cooperative agreement between SACU and US

(Source: the dti, 2014b)

23

Table 2: Percentage Ad Valorem Rate, Import Taxes

South Africa

(i)

Other TFTA countries

(ii)

Grains and Crops 4.178 2.340

Livestock and Meat

Products 2.887 0.006

Coal Extraction 0.934 0.000

Oil and Gas Extraction 2.523 0.000

Other Extraction 0.797 0.002

Processed Food 8.561 1.173

Textiles and Clothing 5.000 0.816

Light and Heavy

Manufacturing 4.752 0.173

Petroleum and Coal

Products 5.077 0.611

Electricity and Gas 0.459 0.000

Total 35.168 5.121

(Source: GTAP database)

Table 3: Aggregated Commodities

Grains and Crops Light and Heavy Manufacturing

Livestock and Meat Products Petroleum and Coal Products

Coal Extraction Electricity and Gas

Oil and Gas Extraction Water

Other Extraction Construction

Processed Food Transport and Communication

Textiles and Clothing Other services

24

Table 4: Back-of-the-Envelope Model

𝐺𝐷𝑃(𝑟) = 𝐶(𝑟) + 𝐼(𝑟) + 𝐺(𝑟) + (𝑋(𝑟) − 𝑀(𝑟)) (1)

𝑀(𝑟) = 𝐹𝑀(𝐺𝐷𝑃(𝑟), 𝑇𝑂𝑇(𝑟),1

1 + 𝑇) (2)

𝑋(𝑟) = 𝐹𝑋(−𝑇𝑂𝑇(𝑟) ∗ 𝐺𝐷𝑃) (3)

𝑇𝑂𝑇(𝑟) = 𝑃(𝑟)𝑋

𝑃(𝑟)𝑀

(4)

𝑅𝑂𝑅(𝑟) = 𝐹𝑅𝑂𝑅(𝐾

𝐿(𝑟), 𝑇𝑂𝑇(𝑟), 𝐴) (5)

Table 5: Macroeconomic Results

Percentage change

Value of

GDP

Volume of

merchandise

exports

Volume of

merchandise

imports

Terms of

trade

Output of

capital goods

sector

Market price

index for

primary

factors

NAFTA -0.005 0.005 -0.006 -0.002 -0.007 -0.005

EU25 -0.005 0.003 -0.005 -0.002 -0.009 -0.005

BRIC -0.008 -0.002 -0.011 -0.004 -0.006 -0.007

ZAF 1.023 0.685 2.003 0.698 1.155 1.004

Other TFTA -0.291 0.547 0.603 -0.162 0.357 -0.023

RoA -0.006 0.008 -0.007 -0.007 -0.015 -0.006

RoW -0.006 0.001 -0.007 -0.003 -0.009 -0.005

(Source: Simulation Results)

25

Table 6: Welfare Decomposition

Million US$

Welfare

Allocative

Efficiency

Effect

Terms of Trade

Effect

Investment-

Savings Effect

Trade Tax

Effect

NAFTA -57.781 -1.725 -33.330 -22.728 -7.166

EU25 -126.053 -10.130 -109.545 -6.380 -59.784

BRIC -127.100 -43.401 -89.747 6.045 -34.419

ZAF 815.985 193.853 602.899 19.930 104.769

Other

TFTA -324.989 -102.814 -221.484 -0.854 15.140

RoA -24.898 -5.835 -18.437 -0.633 -1.981

RoW -154.360 -27.181 -131.742 4.560 -30.321

Total 0.805 2.768 -1.386 -0.060 -13.760

(Source: Simulation Results)

Table 7: Selected Macroeconomic Indicators for South Africa1

Percentage change

GDP quantity index 0.068

Value of GDP 1.023

Private consumption expenditure 1.082

Government consumption expenditure 1.120

Volume of merchandise imports2 2.003

Volume of merchandise exports3 0.685

Change in trade balance, X - M (US$ million) -596.769

Terms of trade 0.698

Output of capital goods sector 1.155

GDP price index 0.954

Price index of merchandise imports -0.012

Price index of merchandise exports 0.686

Rental rate on capital 1.022

Real return to capital 0.291

Real return to unskilled labour 0.314

Real return to skilled labour 0.338

Market price index for primary factors 1.004

Current net rate of return on capital stock 0.685

1 In line with the standard model’s policy closure, the world price index of primary factors is exogenous and acts as a numeraire. 2 Value merchandise imports increase with 1.991 percent. 3 Value of merchandise exports increase with 1.375 percent.

26

(Source: Simulation Results)

Table 8: Selected Industry Results

Percentage change

Value added

Aggregate

exports,

FOB

weights

Aggregate

imports, CIF

weights

Aggregate

exports price

index

Aggregate

imports

price index

Ratio of

domestic to

imported

prices

Grains and

Crops -0.036 -0.418 3.037 0.792 -0.387 1.182

Livestock and

Meat Products -0.045 -1.496 2.565 0.771 -0.035 0.807

Coal

Extraction -0.320 -0.883 0.721 0.207 0.021 0.186

Oil and Gas

Extraction -0.289 -3.314 2.050 0.250 -0.01 0.26

Other

Extraction -0.663 -0.868 0.755 0.491 -0.024 0.515

Processed

Food 1.211 9.771 2.044 0.799 -0.056 0.855

Textiles and

Clothing 0.218 7.940 2.724 0.784 -0.073 0.857

Light and

Heavy

Manufacturing

-0.240 0.694 2.087 0.725 -0.016 0.741

Petroleum and

Coal Products 1.944 10.272 0.477 0.090 -0.011 0.101

Electricity and

Gas -0.336 -2.230 1.683 0.623 -0.065 0.689

Water 0.097 -4.579 2.512 0.837 -0.007 0.845

Construction 0.976 -2.976 2.543 0.787 -0.008 0.795

Transport and

Communicatio

n

-0.170 -3.081 1.588 0.833 -0.007 0.839

Other services 0.065 -3.470 1.953 0.925 -0.006 0.930

(Source: Simulation Results)

27

Table 9: Demand for Endowment for Use in Industries

Percentage change

Land

Unskilled

labour

Skilled

labour Capital

Natural

resources

Grains and

Crops -0.004 -0.045 -0.052 -0.039 0.003

Livestock and

Meat Products 0.012 -0.058 -0.07 -0.046 0.003

Coal

Extraction -0.465 -0.586 -0.591 -0.582 0.000

Oil and Gas

Extraction -0.406 -0.517 -0.521 -0.512 0.001

Other

Extraction -0.630 -0.785 -0.789 -0.78 -0.001

Processed

Food 0.642 1.200 1.172 1.226 0.004

Textiles and

Clothing 0.177 0.215 0.185 0.245 0.003

Light and

Heavy

Manufacturing

-0.031 -0.254 -0.285 -0.225 0.003

Petroleum and

Coal Products 0.931 1.925 1.895 1.955 0.005

Electricity and

Gas -0.074 -0.351 -0.381 -0.321 0.003

Water 0.120 0.088 0.057 0.117 0.003

Construction 0.487 0.965 0.931 0.998 0.004

Transport and

Communicatio

n

0.020 -0.187 -0.226 -0.149 0.003

Other services 0.108 0.061 0.030 0.090 0.003

(Source: Simulation Results)

28

Figure 1: South African Trade as a Percentage of GDP (1960-2013)

(Source: World Development Indicators)

Figure 2: Goods Imported from Other TFTA Countries into South Africa

(Source: GTAP database)

Figure 3: South African Goods Exported to Other TFTA Countries

20

30

40

50

60

70

80

1960

1962

1964

1966

1968

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

2012

Tra

de

(% o

f G

DP

)

29

(Source: GTAP database)

Figure 4: Economic Relationships in the GTAP model

(Source: Hertel & Tsigas, 1997)

Figure 5: Nested Production Structure

30

(Source: Hertel & Tsigas, 1997)

Figure 6: Private and Government Demand for Domestic and Imported Goods

(Source: Simulation Results)

Figure 7: South African Export Sales to Other Regions

-2

-1

-1

0

1

1

2

2

3

3

4

Per

cen

tag

e ch

ang

e

Industry (goods)

Private household demand for domestic goods Private household demand for imported goods

Government demand for domestic goods Government demand for imported goods

31

(Source: Simulation Results)

-7

-5

-3

-1

1

3

5

7

9

11

13

15

17

19

21

23

25

27

29

31

33

35

37

NAFTA EU25 BRIC TFTA RoA RoW

Per

cen

tag

e ch

ang

e GrainsCrops

MeatLstk

Coal

OilGas

OthExtract

ProcFood

TextWapp

Manufactures

PetroRef

ElecGas

Water

Construction

TransComm

OthServices

32

APPENDIX A: The Data

Table A1: Aggregation of Regions

Model Region Regions Included

South Africa ZAF South Africa

Tripartite Free

Trade Area

(excluding South

Africa)

Other

TFTA

Egypt; South Central Africa; Ethiopia; Kenya; Madagascar; Malawi; Mauritius;

Mozambique; Rwanda; Tanzania; Uganda; Zambia; Zimbabwe; Botswana; Namibia;

Rest of South Africa Customs

European Union EU25

Austria; Belgium; Cyprus; Czech Republic; Denmark; Estonia; Finland; France;

Germany; Greece; Hungary; Ireland; Italy; Latvia; Lithuania; Luxembourg; Malta;

Netherlands; Poland; Portugal; Slovakia; Slovenia; Spain; Sweden; United Kingdom

Major emerging

economies BRIC Brazil; Russian Federation; India; China

North America NAFT

A Canada; United States of America; Mexico; Rest of North America

Rest of Africa RoA

Rest of Eastern Africa; Morocco; Tunisia; Rest of North Africa; Benin; Burkina Faso;

Cameroon; Cote d’Ivoire; Ghana; Guinea; Nigeria; Senegal; Togo; Rest of Western

Africa; Central Africa

Rest of World RoW

Australia; New Zealand; Rest of Oceania; Hong Kong; Japan; Korea; Mongolia;

Taiwan; Rest of East Asia; Cambodia; Indonesia; Lao People’s Republic Democratic

Republic; Malaysia; Philippines; Singapore; Thailand; Vietnam; Rest of Southeast

Asia; Bangladesh; Nepal; Pakistan; Sri Lanka; Rest of South Asia; Argentina; Bolivia;

Chile; Columbia; Ecuador; Paraguay; Peru, Uruguay; Venezuela; Rest of South

America; Costa Rica; Guatemala; Honduras; Nicaragua; Panama; El Salvador; Rest

of Central America; Caribbean; Switzerland; Norway; Rest of EFTA; Albania;

Bulgaria; Belarus; Croatia; Romania; Ukraine; Rest of Eastern Europe; Rest of

Europe; Kazakhstan; Kyrgyzstan; Rest of former Soviet Union; Armenia; Azerbaijan;

Georgia; Bahrain; Islamic Republic of Iran; Israel; Kuwait; Oman; Qatar; Saudi

Arabia; Turkey; United Arab Emirates; Rest of Western Asia

33

Table A2: Aggregation of Commodities

Sector Description Commodities Included in Sector

Primary industries

GrainsCrops Grains and Crops Paddy rice; Wheat; Cereal grains; Vegetable, fruit, nuts; Oil seeds;

Sugar cane, sugar beet; Plant-based fibres; Crops; Processed rice

MeatLstk Livestock and

Meat Products

Cattle, sheep, goats, horses; Animal products; Raw milk; Wool,

silk-worm cocoons; Meat: cattle, sheep, goats, horse; Meat

products

Coal Coal Extraction Coal

OilGas Oil and Gas

Extraction Oil; Gas

OthExtract Other Extraction Minerals; Forestry; Fishing

Secondary industries

ProcFood Processed Food Vegetable oils and fats; Dairy products; Sugar; Food products;

Beverages and tobacco products

TextWapp Textiles and

Clothing Textiles ; Wearing apparel

Manufactures Light and Heavy

Manufacturing

Leather products; Wood products; Paper products, publishing;

Chemical, rubber, plastic products; Mineral products; Ferrous

metals; Metals; Metal products; Motor vehicles and parts;

Transport equipment; Electronic equipment; Machinery and

equipment; Manufactures

PetroRef Petroleum and

Coal Products Petroleum, coal products

ElecGas Electricity and

Gas Electricity; Gas manufacture, distribution

Water Water Water

Construction Construction Construction

Tertiary industries

TransComm Transport and

Communication Trade; Transport; Sea transport; Air transport; Communication

OthServices Other services Financial services; Insurance; Business services; Recreation and

other services; Public administration, defence, health and

education; Dwellings

34

APPENDIX B: Results

Table B1: Change in Volume of Total Exports from r to s

(s)

NAFTA EU25 BRIC ZAF

Other

TFTA RoA RoW Total

(r)

NAFTA 87.322 46.352 16.918 140.015 -271.336 5.041 71.529 95.840

EU25 38.065 276.048 24.770 580.401 -977.473 29.043 110.654 81.508

BRIC 67.829 85.405 17.715 225.917 -583.407 11.394 133.200 -41.947

ZAF -492.885 -986.001 -334.793 0.000 3374.466 -110.658 -921.54 528.588

Other TFTA 99.055 247.995 55.109 205.419 -94.312 38.148 184.663 736.077

RoA -4.255 -2.531 -6.361 69.322 -36.218 0.619 -1.533 19.043

RoW 33.433 73.027 39.941 475.909 -745.063 13.756 121.953 12.956

Total -171.437 -259.705 -186.702 1696.983 666.657 -12.659 -301.075 1432.064

(Source: Simulation Results)

35